SE 5879 Oscillation at Full Volume

[RWood]

This is my latest version of what Firemedic built here https://el34world.com/Forum/index.php?topic=11713.0

I like the tone of these amps and this one sounds very good. . . . . . . until the volume is on 9 or 10. Opening up the volume pot sends something into oscillation. When it happens, the signal drops to barely audible, and is distorted. The current draw drops by 5mA and the B+ increases by 5 volts.

The Treble control does influence when the oscillation begins. The higher the treble is set, the lower on the volume pot the oscillation will begin (closer to 9). As treble is reduced that onset gets closer to 10 and when the treble is at it lowest, the volume can be maxed and there is no oscillation.  And no treble.

I feel like there is something basic that I left out of the build, and/or a layout problem but danged if I can find it. This amp only has the one Vol control, unlike Firemedic's, and it doesn't have any switches. It is point-to-point (which looks even worse in 2D). I have used shielded wires for the grids. I've tried 12DW7, 12AT7 and 12AU7 input tubes with only slight differences in when the oscillation begins. Chopsticking hasn't revealed anything.

Any ideas on how to troubleshoot this? Here is the schematic and I will put photos in the next post. Thanks for looking.
 

[RWood]

Pics

[RWood]

a few more pics

[tubeswell]

Didn't have time to read all the wayyy through your other build thread, but, did you put an NFB loop in by any chance?

[tubenit]

until the volume is on 9 or 10

Does this mean 9 or 10 like 9 or 10 o'clock  such as on a wristwatch?


OR does this mean 9 or 10 like   1-10 on a volume pot?   


IF it's this latter one, then don't it play it with the volume at 9 or 10.  In fact, I've never played any amp I've owned with the volume set at 9 or 10. In fact, I can't remember even "testing" to see what any one of my current 3 amps would sound like dialed in that way?

I would say it's rare for me to have the volume set higher than 12 noon or "5". 

I think with the point to point wiring in a small chassis, I'd just accept it as it is and work to enjoy it.  The layout & lead dress likely will contribute to some oscillation.

With respect,  Tubenit

[RWood]

I mean 9 or 10 on the volume control.


Tubeswell - no feedback (yet). That is something I wanted to try before this problem presented itself.

[tubenit]

Try changing the cathode cap on the 5879 to a 1uf or 2.2uf. I think that might help? 


with respect, Tubenit

[pdf64]

My view is that an amp should be unconditionally stable.
I think it’s oscillating because there’s too much noninverting gain between the input stage grid and the DCCF’s grid or cathode.
The tiny bit of capacitance between the 12AU7 sections must be enough coupling to meet the necessary conditions for oscillation.
To get it stable, need to shift the phase / lower the gain, up in the ultrasonics.
Try an anode to cathode cap on the 5879, eg 47 or 100pF.

[RWood]

Pete - it worked! Thank you for the idea. I used a 100p cap.


Tubenit - I would have tried the screen bypass cap first (it was easier to get to) but didn't have a cap with a high enough voltage rating. Thanks for chiming in. Oops I reread your post and you were talking about the cathode bypass, not the screen. I could have done that..... Oh Well.


I am stoked. I like the 5879 tone so much and want to try it on more amps. You guys are awesome!


RWood

[RWood]

I am back with another version of this amp. This one is a combo with a separate control panel, amp in the bottom of the cabinet. I used an eyelet board this time, and there's no oscillation (yay). I also changed the tone stack to Ampeg style, and I am making it bypass-able via a push/pull pot on the bass control. So far I like it a lot.


I would like to tweak a few things though, the first thing being to add some feedback, preferably switchable. I am slightly stumped because the triode I would normally add it to is a cathode follower in this amp. So... would I add to the cathode of the 5879? Suggestions sought here.


Also as you will see I have a spot on the control panel I will need to fill, so a rotating switch for the feedback is kinda what I am seeing there.


And any other suggestions on the schematic would be welcome. Here is the latest version (2) and some pics of the amp in situ will follow.


thanks!

[RWood]

Pics

[RWood]

hey its a sideways combo